Department of Labor Home Connecticut Labor Market Information Home Connecticut Labor Market Information
Home  About  Contacts  FAQ  Glossary  Sitemap  Search  
LMI Calendar   
Connecticut Economic Digest: November 2005 issue

JOBS AND CYCLES: Historical Patterns in Connecticut's Employment Behavior
By Daniel W. Kennedy, Ph.D., Senior Economist, DOL

In keeping with its historical pattern, Connecticut lagged the U.S. in the current jobs recovery. And, as after the 1990-91 recession, job growth has been slow. Connecticut's current recovery has been underway for 23 months. More than 33,000 net, new jobs have been added between September 2003 and August 2005 (the latest data at the time of writing). That represents a 2.01% growth, or a 1.05% compounded, annualized growth-rate. Once again, however, Connecticut lags behind the U.S. Since the U.S. jobs recovery began 27 months ago, U.S. employment has grown by 3.21%, which represents a compounded, annualized rate of 1.42%. Once again, Connecticut led the U.S. in the employment downturn after the bubble burst in 2000, lagged the U.S. in the recent jobs recovery, and has had weaker job growth.

The aim here is not to rehash what has already been written, said, or thought about this problem. Two earlier articles in this publication briefly summarized the history of Connecticut's employment cycle. What follows is based on new research and newly available data that has shed some new light on Connecticut's jobs conundrum. In addition, this analysis also compares the Connecticut and U.S. employment cycles. Newly available data reveals the churning below the surface that exposes the, heretofore, unobserved job flows that produce the net outcomes observed as employment cycles. First, some issues not addressed in this paper should be mentioned before proceeding.

The most obvious difference between Connecticut's job growth and the Nation's lies, of course, in regional differences. Since World War II, the population of the South and Southwest has been growing rapidly, while the North and Northeast have been growing slowly, at best. For 'boom towns' or boom regions and states, trend forces dominate, thus cyclical forces may be minimally felt (although, they are by no means immune from the effects of a severe national recession). The focus here is on the diverging dynamics between the Connecticut and National economic bases. Save the discussion on job flows, which uses quarterly employment data from unemployment insurance program reports, the employment series used in this analysis is the monthly nonfarm employment survey series. Finally, this article is a summary of a longer study that will soon be available on the CTDOL website.

Some Preliminaries. There are two ways to measure a complete cycle: peak-to-peak, or trough-to-trough1. Since the current recovery has not peaked, the trough-to-trough measurement of the complete cycles was used for the following analysis. Finally, three major, distinct periods emerge over the 1945-2003 era. The first, the Post World War II-Cold War Period, began with the end of World War II and lasted until the de facto collapse of the Bretton Woods System in 1971. The Post Bretton Woods-Cold War Period ends with the fall of the Berlin Wall in 1989. The current, Post Cold War Period is dated from the fall of the Berlin Wall in 1989 to 2003. This partitioning will be used to provide the context for what follows.

The Typical Cycle. The following summarizes the 'typical' durations for the various phases of the average employment cycle for Connecticut and the U.S. over the Post World War II era (1945-2003). The average recovery phase for Connecticut was 23.2 months, twice as long as the 11.0 months the U.S. economy spent recovering jobs lost in the recessions. In addition, in the 'typical' cycle since World War II, the U.S. economy had 45.8 months of expansion (i.e., adding net, new employment), compared to only 27.2 months for Connecticut. Connecticut's average recession was 19.1 months in duration, compared to 11.9 for the U.S. Interestingly, while the U.S. typically experienced one-half month of contraction, Connecticut experienced no contraction phase over its employment cycles.

Two Steps Forward, Two Steps Backward. For the entire 58-year span, Connecticut averaged 2.6 months in the recovery/expansion phases for every month spent in the recession/contraction phases, compared to 4.6 months for the U.S. However, though Connecticut has consistently had shorter expansions than the U.S., it did not always lag in job creation. In the Post World War II-Cold War Period (October 1945 to August 1971), Connecticut, on average, added 25.2% to its job base over the recovery and expansion phases, compared to 22.6% for the U.S. This period included the Korean War and the Vietnam War, both stimulants to Connecticut's defense-based economy.

The 1970's brought the collapse of the Bretton Woods System, oil embargos, cost-push inflation, worldwide recession, and challenges to U.S. competitiveness. Consequently, both the U.S. and Connecticut economies faltered. However, Connecticut's job creation ability suffered a bigger setback than the U.S. Consequently, long-term job growth declined more for Connecticut than it did for the U.S. This relative decline in Connecticut's job creation during recovery/expansions resulted in a lower rate of job creation during the average recovery/expansion for the entire Post World War II Era. Connecticut's average job-gain for a 'typical' recovery/expansion phases was 19.4%, compared to 21.1% for the U.S.

However, critical for the State's job-creation performance has been the tendency for Connecticut's economy to, in each recession, take back more of the jobs gained in the previous recovery/expansion phases than the U.S. Though Connecticut's recovery/expansion phases were stronger than those for the U.S. in the Post World War II-Cold War period (1945-1971), its employment recessions were steeper. Over this period, on average, Connecticut's employment base would contract by 5.7% during a recession, compared to only 3.6% for the U.S. From 1971 through the new century, covering the Post Bretton Woods and Post Cold War periods, the development of two mutually reinforcing trends in Connecticut's employment cycle significantly impacted the State's ability to create net, new job growth. Steeper contractions in Connecticut's employment in each downturn, compared to the U.S., exacerbated over the Post Bretton Woods period, and then abated somewhat over the Post Cold War period, but downturns still remained steeper than U.S. employment contractions. The result was that for the entire Post World War II Era, in the average employment downturn, Connecticut shed 5.6% of its jobs, compared to 2.9% for the U.S. In addition, Connecticut's recovery and expansion phases became progressively weaker than those of the U.S.

Divergent Paths. At first glance, it would appear that Connecticut has surpassed the U.S. in transforming into the so-called 'Service Economy'. Based on annual data since 1994, Connecticut's share of service-providing employment surpassed that of the U.S. However, different dynamics can produce the same, observed outcome. And, paradoxically, Connecticut's employment growth displays much more volatility than does the U.S., despite its larger share of service employment. The divergence in their behavior has increased progressively throughout the Post World War II Era, even after Connecticut's share of service-providing employment surpassed that of the U.S. The flip side of the smaller service-providing sector for the U.S. is a larger share of goods-producing employment.

However, a larger share of U.S. goods-producing employment is in construction and mining. Consequently, Connecticut's share of manufacturing employment remains much higher than that for the U.S., even after the hemorrhaging of manufacturing jobs in Connecticut, following the end of the Cold War. In 1990, 83% of Connecticut's goods-producing employment was in manufacturing, compared to 75% for the U.S. By 2004, Connecticut still had three-quarters of its goods-producing employment in manufacturing, compared to two-thirds for the U.S. Further, Connecticut's share of durable goods manufacturing employment has been consistently higher over the 14-year period. In 2004, 9.0% of Connecticut's total nonfarm employment was in the durable goods sector of manufacturing employment, compared to 6.8% for the U.S. That is significant because the variance in the monthly growth rate in durable goods employment is twice that of non-durable goods and seven times that of total nonfarm employment, which contributes considerably to the volatility in Connecticut's employment cycle. Further, Connecticut's service-providing employment contracted by 0.3% between 2000 and 2004 (on an annual basis), while it expanded by 2.3% in the U.S. economy.

Tug-of-War. A new piece to the puzzle comes from newly available job flow data, which provide a glimpse into the churning below the surface of the private segment of the State's economy. Since job creation and destruction data for Connecticut only goes back to 19922, what follows is restricted to exploring the job creation and job destruction data over Connecticut's post-Cold War employment cycle.

Job creation and destruction behavior for U.S. and Connecticut are clearly quite different. These differences in the Job Creation Rate (JCR) and Job Destruction Rate (JDR)3 over the post Cold War employment cycle for Connecticut and the U.S. indicate different underlying dynamics for the two economies. There are two spikes in the Connecticut JCR, at1996:Q4 and 1997:Q44, as well as a sudden drop in job creation in 2003:Q1 (not shown). For the Connecticut JDR, there are two spikes: 1997:Q4 and 2001:Q1. The spikes in the U.S. JCR (down) and JDR (up) during the 2001 recession are what would be expected, but the 1996:Q4 and 1997:Q4 spikes in the Connecticut JCR do not appear in the U.S. JCR.

Also, unlike the U.S., the Connecticut JCR and JDR track each other very closely. Thus, the slightest disturbance can upset the balance between the JCR and JDR and send the Net Employment Growth (NEG) rate into negative territory. The U.S. economy has more of a cushion. Save the jobless recovery at the initial period, and the recession and second jobless recovery toward the terminal period of the cycle, the JCR path is considerably higher than the JDR path for the U.S. The JCR for the U.S. averaged 7.95%, on a quarterly basis, and the JDR averaged 7.56%, for an average NEG of 0.38%. For Connecticut, the JCR averaged 6.80%, and the JDR averaged 6.66%, resulting in a 0.14% average NEG over the cycle. This result provides some new insight into the dynamics driving Connecticut's muted job growth performance, compared to the U.S.

The Next Step. Future research will track job creation and destruction down to the industry level to trace the specific sources of the more intense churning that characterizes the State's economy and contributes to the volatile nature of its employment cycles and the anemic job creation performance. Further, the causes of the anomalous spikes in the fourth quarters of 1996 and 1997 (phenomena not observed in the U.S. data) will be investigated, as they appear to play a significant role in Connecticut's job creation pattern over the post Cold War cycle.

1 For a detailed explanation of the components of the cycle, see Kennedy, Daniel W., .The Connecticut Business Cycle: A Short History (1939 - 2002), CONNECTICUT ECONOMIC DIGEST (June 2002).

2 The U.S. data also go back to 1992, which means that the entire U.S. post-Cold War employment cycle cannot be compared, as its initial trough is in the second quarter of 1991. Therefore, the U.S. data must be compared over Connecticut's cycle definition, which could influence the results.

3 It should be noted that rate has a special meaning here that departs from the conventional definition. This issue is addressed in Appendix A in the complete report.

4 The implications of these spikes are discussed in the complete report.

Return to Top

LMI Home   Contact Office of Research   Site Map   Search 
Published by the Connecticut Department of Labor, Office of Research
Last Updated: November 2, 2005